metadata

dataset_info:
  features:
    - name: audio
      dtype: audio
    - name: sampling_rate
      dtype: int64
    - name: transcription
      dtype: string
    - name: duration
      dtype: float64
  splits:
    - name: train
      num_bytes: 35226517823.384
      num_examples: 136736
  download_size: 33938957274
  dataset_size: 35226517823.384
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/train-*
language:
  - uk
task_categories:
  - automatic-speech-recognition
tags:
  - audio
size_categories:
  - 100K<n<1M

Broadcast Speech Dataset for Ukrainian

Community

Discord: https://bit.ly/discord-uds
Speech Recognition: https://t.me/speech_recognition_uk
Speech Synthesis: https://t.me/speech_synthesis_uk

Statistics

Total duration: 300.181 hours
Number of unique transcriptions: 127441

Duration statistics

Metrics	Value
mean	7.903199
std	3.615765
min	4.99781
25%	5.64
50%	6.65
75%	8.66
max	29.99006

Cite this work

@misc {smoliakov_2025,
    author       = { {Smoliakov} },
    title        = { broadcast-speech-uk (Revision 108837d) },
    year         = 2025,
    url          = { https://huggingface.co/datasets/Yehor/broadcast-speech-uk },
    doi          = { 10.57967/hf/4580 },
    publisher    = { Hugging Face }
}